The ability to produce myriad movements depends on generating rich patterns of muscle activity. Each muscle is driven by the coordinated activity of dozens to thousands of motor units (MUs), defined singularly as a spinal motor neuron and the unique set of muscle fibers it controls. It is the persisting belief of nearly a century that the process by which MUs drive muscles is highly stereotyped and immutable, due to minimal oversight by the brain. However, this perspective primarily derives from studies of steady force production ? a small subset of the kinds of muscular forces that biological creatures produce on a regular basis. The proposed research will accomplish two goals: provide new insights into the process by which MUs generate rich patterns of muscle activity and elucidate the extent of motor cortical involvement in this process. The first two aims will address the first goal, while the second will be addressed by a third aim. Aim 1 will establish whether the central nervous system should have any reason to exert flexible control over individual MUs. This will rely on developing a theoretical model of muscular force production by a collection of MUs and predicting their ideal behavior through mathematical optimization. Aim 2 will investigate whether the nervous system ever does flexibly control different MUs. This aim will involve recording and analyzing muscle activity from the anterior deltoid of rhesus macaques that have been trained to produce the same variety of muscle forces that will be considered in the model (e.g. static, ramp, sinusoidal, chirp). Finally, aim 3 will probe the capacity for primary motor cortex (M1) to exert fine- grained control over single MUs. The conventional wisdom presumes that motor cortex cannot regulate the activity of individual MUs and instead relies on spinal mechanisms to convert coarse-grained descending signals into MU activations. This claim will be investigated by analyzing the responses of MUs to microstimulation in M1, which has recently been shown to possess direct (monosynaptic) projections to MUs that control muscles of the hand, elbow, and shoulder. The proposed research will provide the first theoretical basis for understanding how MUs are used to produce wildly different patterns of muscle activity, the first empirical report of MU activity during voluntary production of rich patterns of muscular forces, and the first investigation into the functional relevance of the direct pathways between M1 and spinal motor neurons. Consequently, this work will shed new light on the normal function of motor neurons, which could profoundly shift our understanding of the cortico-spinal circuits that underlie skilled movements and lead to better treatment for motor neuron diseases such as ALS.